[Paper] Does Self-Evaluation Enable Wireheading in Language Models?
Published on December 8, 2025 4:03 PM GMTTL;DR: We formalized and empirically demonstrated wireheading in…
Recent Articles
View AllAlgorithmic thermodynamics and three types of optimization
Published on December 8, 2025 3:40 PM GMTContext:I (Daniel C) have been working with Aram…
From Barriers to Alignment to the First Formal Corrigibility Guarantees
Published on December 8, 2025 12:31 PM GMTThis post summarizes two related papers that will…
Top 5 Agentic AI LLM Models
In 2025, “using AI” no longer just means chatting with a model, and you’ve probably…
Hinge’s new AI feature helps daters move beyond boring small talk
Hinge launched “Convo Starters,” a feature powered by AI that provides personalized tips for initiating…
OpenAI boasts enterprise win days after internal ‘code red’ on Google threat
ChatGPT Enterprise usage grew 8x year-over-year as workers report saving an hour daily, but OpenAI…
Transforming Nordic classrooms through responsible AI partnerships
Schools across Northern Europe are safely and responsibly integrating Google and Gemini for Education tools…
Scaling what used not to scale
Published on December 8, 2025 8:40 AM GMTThis is part 2/2 of my introduction to Live…
The effectiveness of systematic thinking
Published on December 8, 2025 8:38 AM GMTThis is part 1/2 of my introduction to…